Antipreneurial behavior
in conflict over norms
A case study on the resistance of nation-states against a
preventive ban on lethal autonomous weapons systems
Felix Sippel
Two-year Political Science MA program in Global Politics and Societal Change Dept. of Global Political Studies
Course: Political Science Master's thesis ST631L (30 credits) Spring Semester 2020
Supervisor: Johan Brännmark Submission date: 25 May 2020
Abstract
Since 2014, the international community has been discussing how to deal with the emergence of increasingly autonomous weapons systems under the auspices of the United Nations Convention on Certain Conventional Weapons. This case study examines the behavior of those nation-states that oppose a preventive ban on lethal autonomous weapons systems in this forum. In this regard, the concept of antipreneurship is applied to the analysis of the international meetings in order to explore the resistance patterns with which antipreneurs reject the need for normative change. Analyzing the content of related documents shows that antipreneurs block formal negotiations on a ban and deem legal weapons reviews to be a sufficient regulatory instrument. Beyond that, the antipreneurs also attempt to create a beneficial image of the relationship between artificial intelligence and international humanitarian law. This behavior contributes to force norm entrepreneurs increasingly into a defensive position and sustainably undermines their entrepreneurial demand for a preventive ban. Besides, the analysis distinguishes the resistance of antipreneurs from ambivalent behavior that cannot be clearly linked neither to the behavior of entrepreneurs nor to that of antipreneurs.
Key words: norm antipreneurs, norm resistance, norm contestation, lethal autonomous
weapons systems, artificial intelligence, international security
List of contents
1 INTRODUCTION ... 5
1.1AIM OF THE THESIS ... 5
1.2ACADEMIC RELEVANCE ... 6
1.3CONTENT STRUCTURE ... 7
2 STATE OF ACADEMIC RESEARCH ON LAWS ... 8
2.1DEFINITION OF THE TERM ... 8
2.2RISKS OF LAWS ... 10
2.3REGULATORY DEBATE ... 12
2.4PATTERNS AND GAPS OF RESEARCH... 14
3 THEORETICAL CONSIDERATIONS ON NORMS ... 16
3.1NORM EMERGENCE AND ENTREPRENEURS ... 16
3.2NORM RESISTANCE AND THE DUAL QUALITY OF NORMS ... 18
3.3CONCEPT AND ROLE OF ANTIPRENEURS ... 20
3.4REFLECTION ON ANTIPRENEURSHIP ... 21
4 RESEARCH DESIGN ... 23
4.1SCOPE OF THE CASE STUDY ... 23
4.2COLLECTION OF THE DATA ... 24
4.3ANALYSIS OF THE DATA ... 26
5 FINDINGS OF THE ANALYSIS ... 29
5.1BLOCKING FORMAL NEGOTIATIONS ON A BAN ... 31
5.1.1 Pattern A: Lack of definitions ... 31
5.1.2 Pattern B: Mandate of the forum... 32
5.2DENYING A PROBLEM BETWEEN LAWS AND IHL ... 34
5.2.1 Pattern C: Principles of IHL... 34
5.2.2 Pattern D: Compliance of LAWS ... 35
5.2.3 Pattern E: Legal weapons reviews ... 36
5.3CREATING A BENEFICIAL IMAGE ABOUT LAWS ... 38
5.3.1 Pattern F: Inappropriate claims ... 38
5.3.2 Pattern G: Advantages of AI ... 38
5.3.3 Pattern H: Benefits for IHL ... 40
5.4THRESHOLDS BETWEEN ANTIPRENEUR AND AMBIGUITY ... 42
5.4.1 China: Proponent of a ban? ... 42
5.4.2 France: Proponent of a compromise?... 43
6 CONCLUDING REMARKS ... 45
6.1REFLECTION UPON THE RESEARCH QUESTIONS ... 45
6.2LIMITATIONS OF THE FINDINGS ... 47
6.3IMPLICATIONS FOR RESEARCH AND SOCIETY ... 48
7 LIST OF REFERENCES ... 51
List of figures
Fig. 1: Main areas of academic research on LAWS ... 14
Fig. 2: Tripartite structure of the new norm proposed by the entrepreneurs ... 17
Fig. 3: Guiding questions of the data analysis ... 27
Fig. 4: Resistance patterns of the six antipreneurs ... 30
Fig. 5: Exclusive frame of antipreneurs on the question of compliance with IHL... 37
Fig. 6: Role spectrum between entrepreneurship and antipreneurship ... 44
1 Introduction
1.1 Aim of the thesis
In 2017, the Stockholm International Peace Research Institute already identified multiple deployed weapons systems with increasingly autonomous targeting capabilities sufficient to engage targets without involvement of a human operator (Boulanin and Verbruggen, 2017). Due to these findings, it becomes clear that the future of warfare on the basis of lethal autonomous weapons systems (LAWS) is no longer purely theoretical. Thereby, research outlines that the transnational nature of military technologies which are based on artificial intelligence (AI) creates pressure for global solutions on their development and use (Erdélyi and Goldsmith, 2018: 96).
In this context, civil society coalitions such as the Campaign to Stop Killer Robots as well as large parts of the research community have been driving forces in urging the international community to introduce a preventive ban on the development, production and use of LAWS. Meanwhile, this demand has also been expressed by the Secretary-General of the United Nations, Antonio Guterres, and by more and more nation-states, especially from the global South, in the Group of Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons Systems (GGE LAWS). This forum has been established under the United Nations Convention on Certain Conventional Weapons (CCW) to consult on these emerging weapons systems. However, especially nation-states, which are considered to be leaders in research and development of LAWS, oppose the call for a preventive ban in this forum.
Consequently, this conflict over norms merits an academic assessment. In this respect, the debate on LAWS reflects an imbalance in the research about norms, which typically focuses on those actors who in the role of the entrepreneur construct the need for a new norm. However, recent research has provided a conceptualization of antipreneurs, who, unlike entrepreneurs, defend the status quo of existing norms (Bloomfield, 2016). Since this phenomenon has not yet been studied in the case of LAWS, this thesis aims to examine in detail the behavior of those nation-states that are opposing a ban on LAWS and to discover the characteristic patterns of their resistance:
Research question 1: How do norm antipreneurs resist the call for a preemptive ban on LAWS at the international meetings of the GGE LAWS?
This is a research question that is urgently needed, since the subject of LAWS is still widely overlooked, not only in studies on antipreneurship, but also in the broader empirical research on norms: “the autonomy of weapons systems is accelerating largely outside of public and academic attention in the discipline of International Relations” (Bode and Huelss, 2018: 393). Moreover, the topic should be of great interest for norm researchers, particularly since the relationship between increasingly autonomous weapons systems and the global norms that exist around international security, warfare and humanitarianism is subject to controversial debate – as will be shown in the course of this thesis. In this context, it is of great relevance to find out to which norms the antipreneurs refer in their resistance against a ban and how they describe the relationship between LAWS and the norms that structure global politics. This thought therefore leads to the additional research question:
Research question 2: To which existing norms do antipreneurs relate the subject of LAWS and how do they utilize these norms in their resistance against a ban?
Consequently, the research questions will serve as the red thread of this thesis to analyze the research puzzle in a problem-oriented way. The line of argumentation is based on the view that the demand for a preventive ban on LAWS can be seen as the emergence of a new global norm promoted by norm entrepreneurs. However, the emergence of this new norm is contested by norm antipreneurs. The analysis explores the resistance patterns of antipreneurs and what role existing norms play in these patterns. This is done through a case study about the international talks on LAWS under the auspices of the CCW. Official statements and working papers will be examined through a close reading that moves between the content of these texts and their context in order to find answers to the research questions and to reflect on already available insights on antipreneurship.
1.2 Academic relevance
The research questions are clearly of relevance for the research on global norms. It is drawn particular attention to the role of resisting actors, who are often not as much under consideration by empirical research as the norm entrepreneurs (Bloomfield, 2016; Wiener, 2014). Nor has much attention been paid so far to linking LAWS with literature on international relations, in general (Bode and Huelss, 2018: 393). Particularly the research field on norm resistance and antipreneurship will be enriched by the analysis of
the primary and already very controversial case to empirically examine the early efforts to negotiate global solutions on the use of (military) artificial intelligence.
This research is also relevant because it draws awareness to far-reaching consequences of weapons based on artificial intelligence. Even on more modest readings, it is argued that these technologies may offer nation-states critical advantages across strategic and military domains (Dafoe, 2018; Payne, 2018). Consequently, Maas (2019: 286) states that “there is today a widespread perception amongst the public, policymakers, and scholars, that the global development of AI is swiftly escalating into a strategic arms race — or even a new Cold War — between major states such as the United States and China.” This quotation shows the high actuality and importance of the research puzzle for the present and future of international security and arms control. To understand the conflict over its regulation in the GGE LAWS is thereby a crucial asset. Ultimately, this thesis identifies and addresses a controversial gap in research about norms and is working on one of the most important challenges for security policy in the 21st century.
1.3 Content structure
The further thesis is structured in five additional chapters which are building up on each other. In order to scrutinize the research questions presented above, it is necessary to provide a common understanding of the frequently used terminologies of this work and to localize the concepts in the current, academic debate. So, the next chapter sets the scene by showing a review of academic literature dealing with the matter of LAWS. Patterns of research as well as existing gaps and controversies will be identified. Based on this overview, the subject will be linked with theoretical models which are provided by academic literature on the emergence and resistance of norms in chapter 3.
As soon as the theoretical fundament has been placed, chapter 4 introduces the research design of this work, by which the research questions are operationalized. Thereby, the case study approach will make practical use of the earlier presented academic insights. Next will be the core of the work, as chapter 5 provides the analysis of the case based on the analytical approach. This chapter presents the findings of the analysis in four steps, ranging from resistance patterns that block negotiations on a ban, over the denial of a problem between LAWS and current international norms and the creation of a beneficial image, to the thresholds of antipreneurship. Consequently, chapter 6 discusses the conclusions that can be drawn from the analysis – with particular emphasis on the limitations of the findings as well as on its contributions for society and future research.
2 State of academic research on LAWS
Academic literature on LAWS originates from a wide range of disciplines, including computer science, robotics, law, philosophy and social sciences. The research can be roughly grouped around the following subjects, which are all of relevance in regard of the conflict over a ban on LAWS: first, the question of defining the term. Second, the potential risks which might lead to arguments for a restriction. And third, the debate about how LAWS may be restricted or governed. The following outlines the research patterns in relation to these academic discussions.
2.1 Definition of the term
The significance of the debate on the definition lies in the fact that all its components need clarification. Typically, the term weapon system is described as the ensemble of one or more weapons and a mechanism for firing the weapon, accompanied by a weapon platform, if it is necessary for the system (Egeland, 2016: 92-93). Some authors criticize the anthropomorphic focus of the debate on individual weapons and formulations such as "killer robots" by civil society organizations, which may contribute to a misconception of humanized weapons systems (Roff and Danks, 2018; Van Rompaey, 2019). Instead, it is argued that the discussion should rather focus on the capability of collaborating weapons systems, while these authors refer to network-based concepts (Liu, 2019), such as the Internet of Battle Things (Kott et al., 2016).
The biggest dispute is developing over the addition of the next term to the definition, since the innovation introduced by LAWS is the autonomous agency of such weapons systems and not the system itself. This new quality arises from the application of AI to military purposes (Horowitz, 2019: 767), whereas AI “is not a single technology, but rather a set of techniques and subdisciplines” (Gasser and Almeida, 2017: 59). Research from a technological perspective makes it clear that intelligence of machines could be understood as “that quality that enables an entity to function appropriately and with foresight in its environment” (Dafoe, 2018: 5; Nilsson, 2010). Accordingly, it is the ability of the foresighted application of intelligence that enables these weapons systems to act autonomously.
However, recent research indicates that the concept of autonomy is still not uniformly defined neither in politics nor research. This is because the question arises of how much autonomy and how little human control such weapons systems are meant to have in order
to be regarded as LAWS (Boulanin and Verbruggen, 2017; Van Rompaey, 2019). For instance, a system could have autonomous features but still with complete human control over the use of the weapon. That would be a system with a high degree of automation, although it is not defined as LAWS in most perspectives (Horowitz, 2019: 767; Gill, 2019).
In accordance with this exclusionary approach, Scharre and Horowitz (2015: 16) state that an autonomous weapon system “is intended to select and engage targets where a human has not decided those specific targets are to be engaged.” In contrast, human-supervised weapons systems give human operators the ability to monitor the performance and intervene to halt its operation, if necessary. The definition supported by the US-American government, which is also often quoted, describes it as “a weapon that, once activated, can select and engage targets without further intervention by a human operator” (Boulanin and Verbruggen, 2017: 8).
Thus, these viewpoints rather point towards a narrow definition of LAWS. Nevertheless, these two definitions show a small but significant discrepancy: While the first one mentions that the system is intended for selection and engagement, the US-American one rather formulates its capability for selection and engagement. The latter approach also corresponds with the categorization by Human Rights Watch (2015: 6), which defines autonomous, out-of-the-loop weapons as capable of selecting targets and delivering force
without any human input or interaction. Egeland (2016: 94) adds that it is key that these
fully autonomous weapons systems “have the ability to make decisions themselves, thus breaking the causal chain between humans and the use of force.”
However, Human Rights Watch (2015: 6) usually also includes on-the-loop weapons into the discussion, which allow for the oversight of a human operator, but sometimes with such limited oversight that the weapons are effectively out-of-the-loop. Other political actors such as the Holy See and the International Committee of the Red Cross tend to prefer far more inclusive approaches and overarching concepts (Boulanin and Verbruggen, 2017: 8).
Finally, the term "lethal" is regularly included into the terminology by the research community, but not by all of them. Human rights groups, in particular, tend not to use the term in order to also include non-lethal weapons systems into the debate. The most important example of the addition of the term is the GGE LAWS. However, research indicates that there is confusion about a precise definition even within this forum. LAWS are defined most strictly by delegations such as the French one, only covering systems
that presuppose a complete lack of human supervision and certainly aim at lethal effects (Van Rompaey, 2019: 113).
Finally, the present research, which deals with the question of definition, outlines a trend towards an understanding of LAWS as weapons systems which select and engage targets without human input nor oversight and thus breaking the causal chain between humans and the use of force. However, this understanding is not undisputed, as so far neither governments nor the research community have been able to agree on the exact wording and formulation of a generally accepted definition. This controversial dispute should be carefully considered when preparing the forthcoming analysis, as it may be assumed that it could play a role in the resistance against a ban.
2.2 Risks of LAWS
The previous chapter has already given an insight into how controversial the subject of LAWS can be. The next main area of discussion is also highly relevant to the thesis, since the examined risks of LAWS suggest the arguments for its regulation and even for its prohibition. In this respect, two focal points predominate in the research contributions. First, the geopolitical risk of an arms race is examined. Second, the ethical and legal compatibility of LAWS with current international norms is discussed.
Regarding the first focal point, it is mentioned that AI technology could offer nation-states decisive advantages in many strategic and military domains (Dafoe, 2018; Payne, 2018). In this context, many studies share the view that the application of AI for military purposes will lead to a strategic arms race between major states (Auslin, 2018; Geist, 2016; Lee, 2018; Maas, 2019). Some authors thereby invest in the analysis of China's rising power in the field of military AI and discuss its geopolitical implications and the competition with the USA (Kania, 2017; Lee, 2018; Wyatt, 2020).
Together with China and the USA, other countries are identified with the capabilities of developing LAWS, especially Russia, Israel, South Korea and the United Kingdom (Gill, 2019; Haner and Garcia, 2019). Besides, the military capabilities of countries such as Australia, Canada, France, Germany, India and Japan are also discussed. Another research strategy is to draw comparisons with the history of other technologies, especially nuclear energy and the beginning of the Cold War (Butcher and Beridze, 2019; Garcia 2018; Maas, 2019). The typical view drawn from and supported by these research
activities is that there is a risk of an escalating arms race, which therefore demands the need for international governance.
However, some authors oppose the narrative of an arms race by pointing to backward thinking patterns and the risk of adversarial argumentation, but these contributions are rather a minority (Cave and ÓhÉigeartaigh, 2018; Zwetsloot et al., 2018). It is nevertheless instructive that, although the mention of an arms race is widespread, it is usually associated with hypothetical discussions on how to prevent it. In contrast, research on the extent to which an arms race is already taking place and who is taking what position in it remains rather hesitant and has little empirical foundation.
Considerably more fruitful in terms of its academic foundation is the state of research on the second focal point, for which several legal scholars critically analyze the risks of LAWS and conclude that they are hardly compatible with existing international norms (Davison, 2017; Egeland, 2016; Sharkey, 2017). In particular, it is argued that LAWS contradict two guiding principles of International Humanitarian Law (IHL) – also called the laws of armed conflict – namely the distinction between combatants and civilians and the principle of proportionality (Asaro, 2012; Roff, 2015; Sharkey, 2010; Solovyeva and Hynek, 2018).
The terminological hurdles alone, such as the lack of a clear definition of the term "civilian", would make it difficult to translate such specifications into a computer code (Sharkey, 2017: 179). It is further argued that precisely these difficult decisions in combat – such as the distinction between military targets and civilian objects and the weighing of an attack against its military objective and the possible civil damage – require an interpretation that must be based on human judgement due to “subjective estimates of value and context-specificity” (Heyns, 2013: 13).
This human component is also crucial with regard to another major argument put forward by scholars, since IHL, like other normative instruments, requires human actors to hold someone responsible for possible crimes (Solovyeva and Hynek, 2018: 194). The lack of human supervision could lead to the case that “a weapon system will have either no one or too many people to be held accountable for mistakes” (Garcia, 2015: 60). It is also mentioned that this responsibility gap could even incentivize the further development of LAWS in order to avoid liability for the conduct of war (Grut, 2013; Liu, 2012).
Consequently, these analyses draw attention to the legal conflict potential that lies in the use of LAWS, which may not be consistent with the international principles of distinction, proportionality and responsibility in armed conflict. Moreover, these concerns are accompanied by ethical considerations that the killing of humans by autonomous systems may be considered an immoral scenario and against human rights (Heyns, 2016; Korać,
2018; Rosert and Sauer, 2019; Johnson and Axinn, 2014). The conclusion drawn from
and supported by these research activities is that the application of LAWS puts international norms at risk, which necessitates the need for normative action.
Even scientists who see great potential in LAWS respect the fact that achieving the capacity to reliably comply with international norms would be a necessary precondition for the legitimate use of LAWS (Lark, 2017; Umbrello et al., 2020). However, it is also stressed that technological advances could lead to machines capable of avoiding many human mistakes and being able to “outperform human soldiers with respect to conformance to IHL” (Arkin, 2018: 323). Based on such counterarguments, it is suggested that LAWS will help to make conflicts more ethical by limiting the number of victims, unnecessary destruction and trauma (Arkin, 2010; Arkin, 2015; Lucas, 2014). Moreover, the argument of the responsibility gap is countered on the basis that it would rely upon implausibly idealized understandings of human decision-making in combat (Birnbacher, 2016; Smith, 2019). Besides, it is proposed that LAWS shall remain under the tactical control and under the responsibility of the military commander in the battle (de Boisboissel, 2017).
In summary, it can be said that this controversy reveals two groups of researchers. On the one hand, the scholars just quoted, having faith in technological progress, who want to emphasize the positive potential of LAWS in future scenarios. On the other hand, researchers who are particularly proficient in the field of law refer to existing norms in the context of armed conflict and clearly highlight the humanitarian and ethical risks that exist with regard to the emergence of LAWS.
2.3 Regulatory debate
The two previous sub-chapters have already outlined a broad research debate in which two general trends can be identified: first, the complexity of the subject, to which many scientific perspectives contribute. Second, the uncertainty surrounding the topic, characterized by contradictory viewpoints and some forms of guessing games. Based on the arguments already described, the next chapter outlines academic research on the regulation of LAWS.
In this respect, the area of discussion is usually approached in scientific research by theoretical and normative proposals as to which regulatory concepts are regarded as appropriate and inappropriate, with an international ban on LAWS at the center of the debate. In fact, many scientists argue for this most rigorous form of restriction (Asaro,
2012; Goose, 2015; Rosert and Sauer, 2019; Wallach, 2017). In doing so, they suggest an international treaty (Asaro, 2012; Goose, 2015) and draw comparisons with approaches with regard to other prohibited weapons such as anti-personnel mines and cluster munitions (Rosert and Sauer, 2019).
However, there are also scientists who argue against a preemptive ban, primarily with the arguments that such a step would be too early for the time being and that an international ban would not be effectively controllable due to the discrete nature of AI-related technologies (Arkin, 2015; Danet, 2017; Jenks, 2017).
Apart from this, there is also a broad literature that advocates a certain restriction but does not plead for a complete ban. Some of these alternatives see LAWS as part of an overarching AI framework that needs to be created (Erdélyi and Goldsmith, 2018; Scherer, 2016; Sethu, 2019). Others argue for more problem-specific approaches, suggesting agreements on the applicability of the existing laws of armed conflict (Anderson et al., 2014), a new legal architecture of responsibility (Liu, 2012) or an agreement on the concept of meaningful human control (Cummings, 2019; Sethu, 2019). And few articles focus more on international standards for the design of robots and programming languages that could be the key to effective bottom-up regulation (Cihon 2019; Henshaw et al., 2010; Sharkey, 2017).
Apart from the contributions on how LAWS should be regulated in the future, there is also academic literature that provides empirical research on the (inter-)national efforts that have already been made. For example, there are reports on already existing national frameworks and strategic concepts with regard to AI (Butcher and Beridze, 2019; Simon, 2019). There are also some empirical studies relating more specifically to international talks on LAWS. For instance, Akimoto (2019) thereby examines Japan´s role in the GGE LAWS through different theoretical lenses of international relations, especially neo-liberalist and neo-realist approaches. And Bode and Huelss (2018) do groundbreaking work as they base their research on norms literature and assess how procedural norms might be emanating from state practices with regard to LAWS.
Also, of relevance for this thesis is Danet (2017), who describes the emergence of the international debate by splitting his analysis in the phase of lobbying from civil society actors such as the Campaign to Stop Killer Robots and the phase of institutionalization by initiating the GGE LAWS as the principle forum for international discussions on this matter. Bode (2019) also focuses on this UN-driven forum and thereby on the
norm-making attempts of countries from the global South which are promoting a preventive ban in the GGE LAWS.
What these studies have in common is that they refer to the international debate on LAWS in terms of a case-related analysis of qualitative data, which was mostly collected through official documents and statements, in the case of Bode (2019) also through interviews. Consequently, the methods and findings of these few empirical studies are very helpful in identifying relevant methodological aspects for the upcoming analysis.
2.4 Patterns and gaps of research
As this concludes the literature review, figure 1 illustrates the identified main areas of academic debate and its major patterns. The terminological debate is characterized by the understanding of autonomy and critique of the anthropomorphic focus. The geopolitical debate focuses on the military capabilities of nation-states and on a possible arms race, with historical comparisons being made, especially in terms of nuclear energy. The normative debate examines LAWS in the context of the principles of IHL – particularly the principles of distinction and proportionality – and points to a responsibility gap and ethical concerns. And the regulatory debate revolves around the questions of whether a preemptive ban is necessary, what other regulations would make sense and what can be learned from the current state of discussions in the GGE LAWS.
All these contributions outline the extent to which the research community is involved in the formulation of definitions, risks and regulatory approaches, frequently blurring the line between neutral research and near-political standpoints. It becomes clear that the research community cannot only be seen as an impartial investigator of discourse, but rather as a committed actor itself.
However, this does not have to be considered a negative sign in principle, since the complexity of this subject area even necessitates the active though prudent participation of scientific experts (Righetti et al., 2018: 125). Thereby, Haas (1992: 4) defines them as part of the epistemic community based on “networks of professionals with recognized expertise and competence in a particular domain.”
By outlining patterns and controversies, the research questions can be placed in the context of the research about the international meetings in the GGE LAWS and the discussion about a ban on LAWS. However, it can be assumed that all areas of academic research shown in figure 1 might be central to the forthcoming analysis, since terminological, legal, ethical and geopolitical considerations are likely to play an important role within the conflict over norms, too.
However, the literature review reveals that empirical research on the discussions in the GGE LAWS and the linking of this matter with research about global norms only represents a small and underrepresented piece of the overall research puzzle, since “the autonomy of weapons systems is accelerating largely outside of public and academic attention in the discipline of International Relations” (Bode and Huelss, 2018: 393). The thesis is therefore aimed at filling one of these research gaps, as there has not been any empirical research so far about the resistance against a preventive ban on LAWS in these international meetings. Its theoretical foundation within the literature on norms will be discussed in the following chapter.
3 Theoretical considerations on norms
In the previous chapter, it was possible to identify a small number of initial studies which put the topic in the context of norms. Since it is necessary to create a common understanding of the applied terms of this thesis, this context will be further specified. Thereby, literature on global norms is an important pillar of the research program in international relations. The most common definition sees a norm as “a standard of appropriate behavior for an actor with a given identity” (Finnemore and Sikkink, 1998: 891). Tannenwald (1999: 436) adjusts this definition in her study about efforts to ban nuclear weapons and defines a norm as “a shared expectation about behavior, a standard of right or wrong.”
Based on this understanding of a norm, the question arises how to examine the conflict over a ban on LAWS from a perspective of norms research? This is addressed in the course of this chapter in the following manner: First, concepts of norm emergence and entrepreneurship will be discussed in order to situate the side of the actors who are in favor of a ban. Then, the chapter turns to those actors who oppose such a demand in terms of norm resistance and contestation, whereas the concept of antipreneurship is introduced.
3.1 Norm emergence and entrepreneurs
To begin with, the two definitions introduced above outline the function of a norm to structure the behavior of an actor. By this, norms can shape both the behavior of individuals in everyday life situations as well as the behavior of nation-states and organizations in the context of international relations. In such instances, norms set a framework for international action on issues such as trade, security and human rights while “providing a language and grammar of international politics” (Cortell and Davis, 2000: 65-66).
But how does such shared expectation about behavior emerge? By this, the relationship of norms to entrepreneurship has to be clarified. Typically, academic models such as the frequently used norm life cycle emphasize the role of norm entrepreneurs, who try to convince a critical mass of nation-states to adopt new norms on the international scale. In doing so, they draw attention to problems or even create problems by using interpretive language (Finnemore and Sikkink, 1998: 895-901). Accordingly, the advocates of a preventive ban on LAWS can be seen as norm entrepreneurs. Moreover, the entrepreneurs thus problematize the status quo of current norms and plead for change, since the
emergence of new norms may not take place in a vacuum, but in controversial domains of already existing normative understandings.
A recent model presented by Rosert (2019) outlines this entrepreneurial quest in terms of an empirical case analysis of the ban on cluster munitions. The model takes particular account of norm-setting efforts of social actors from the bottom up, who aim at formalized arrangements. These efforts are similar to the case of LAWS, whereas human rights groups have been the main driving force behind the promotion of a ban before the international talks in the GGE LAWS began. According to the model of Rosert (2019: 1107-1110), the entrepreneurs take up something they perceive as a problem: In terms of LAWS, this problem could be that emerging weapons systems will act autonomously in deciding who to kill. Then, the entrepreneurs try to raise this problem to the public and, more importantly, to place it on the institutional agenda in order to create a new international norm. In this sense, norm entrepreneurs such as Human Rights Watch and the Campaign to Stop Killer Robots already have succeeded in putting the subject of LAWS on the institutional agenda through the GGE LAWS.
Moreover, it is useful to introduce Winston´s (2018: 640) tripartite structure of norms to the discussion, which similar to Rosert (2019) is based on the adoption of a presupposing problem by norm entrepreneurs. This tripartite structure is helpful not only to outline the efforts of the entrepreneurs in creating a new norm but also their reasoning as to why this norm is needed, see figure 2. Thereby, the perceived problem of LAWS may be linked with a value, which is the enjoyment of something good or the avoidance of something bad and as such gives moral weight to the problem.
Based on the discussion outlined in chapter 2.2, these values could be based on the principles of IHL, such as always to distinguish between soldiers and civilians. Based on this linkage, it can be argued that the problem of autonomously killing weapons systems hinders the full enjoyment of these values and therefore requires corrective behavior to enable the actor to better express the values. In the given case, entrepreneurs argue that it shall result in the behavior of nation-states not to develop, produce, deploy or use LAWS (ibid.: 640). This tripartite model thus provides an argumentative structure for the preventive ban demanded by entrepreneurs.
In addition, Wiener (2017: 716) introduces a categorization that outlines the effects of the behavioral claims behind such a norm on the global society. Here, fundamental norms such as the one discussed above, which are negotiated at the macro level of global society by government representatives, are leading to organizing principles that are constituted by policies and political practice at the meso level. Finally, standards and regulations are established by norm followers, such as companies at the micro level of global society. This categorization exemplifies how the behavioral claim not to develop LAWS could be translated and implemented into regulations for programmers and military industry. On the macro layer, such a new norm could be formalized by legally binding instruments such as treaties. These may be the obvious norms that structure global politics (Panke and Petersohn, 2011). It is also possible to think of purely political, non-binding arrangements and other non-codified norms in accordance with customary law (Erdélyi and Goldsmith, 2018: 96). It means that the structure of figure 2 would not necessarily have to be enshrined in a binding instrument. However, the efforts of entrepreneurs tend to focus on the introduction of codified frameworks to assure responsibility and compliance.
Consequently, the models presented so far are suitable for classifying the emergence of norms and the way how they may structure behavior. However, the focus of these models is typically on those actors who, in their role as entrepreneurs, agree on the need for a new norm. In an extension to this, the literature on norms is enriched by scientists who place greater emphasis on complexity and resistance, which will be addressed in the upcoming sub-chapter (Wiener, 2014; Zimmermann, 2016).
3.2 Norm resistance and the dual quality of norms
By addressing the complexity of norms, it opens a perspective on the manifold relationships between norms and actors that exist in addition to entrepreneurship. In this context, research particularly points to the idea that norms have two contradictory
qualities: while the recognition of the existence of a norm structures behavior, the content and validity of norms are constantly contested and reconstituted (Winston, 2018: 642-643). This behavior in turn can influence the content and meaning of the norm in order “to simultaneously structure and be structured by social interaction” (Lantis and Bloomberg, 2018: 165-166). Basically, this perspective can be traced back to other dualities in social sciences, such as the agent-structure and micro-macro approaches, which show that explaining social behavior must reflect contexts and interdependencies (Wendt, 1987: 337-339).
Furthermore, the dual quality of norms implies to add a further perspective on those actors who do not support the existence or emergence of a norm. Such actors can apply a “range of social practices, which discursively express disapproval of norms” that “may be performed either explicitly (by contention, objection, questioning, or deliberation) or implicitly (through neglect, negation, or disregard)” (Wiener, 2014: 1-2). In order to scrutinize these practices, one group of scholars thereby focuses on a bottom-up contestation by non-state actors, such as social movements, which are problematizing the effects of global norms at the local and regional level (Beveridge and Naumann, 2014; Wiener, 2014). In the majority of empirical research, there is usually a focus on already existing norms and how contestants might be able to change their meaning-in-use. However, the case of LAWS refers to the emergence of a new norm. On this account, a close look at figure 2 outlines grounds for contesting the claims of entrepreneurs: there may be different perceptions of the content, scope or implications of the problem; as well as different views on the relationship of a problem to certain values and on the range of behavioral implications resulting from this relationship. Based on these considerations, resisting actors could for instance show an ambiguous behavior in which they agree that there is a problem which requires normative change, but these actors favor to restrict them in a different way that would make a preemptive ban dispensable. Research refers to them sometimes as competing or rival entrepreneurs (Lantis and Wunderlich, 2018: 573-574). In addition, a number of alternative approaches that could be available to such actors have already been mentioned in chapter 2.3.
However, in contrast to Winston (2018) and Rosert (2019), there could also be actors who do not perceive something as a problem at all. By challenging the problematization as a whole, these actors therefore may see no proper reason for a new norm. Here, the focus shifts, since resistance is neither understood as the effort to contest the meaning of already established norms, nor as the effort to address a problem in another manner. Rather, the
focus here is on the effort to reject the claims of entrepreneurs and to resist against any normative change. This will be further conceptualized in the next sub-chapter.
3.3 Concept and role of antipreneurs
The resistance against the arguments of the entrepreneurs and the normative change demanded by them constitutes the role for which the concept of antipreneurship manifests itself. In particular Bloomfield (2016: 321) introduced this concept, which he defines “in polar-opposite terms to entrepreneurs, namely, antipreneurs are actors who defend the entrenched normative status quo against challengers.” He also argues that entrepreneurs during the phases of norm emergence must establish that there is a problem with the status quo of current norms; then they must offer a viable solution in the form of a new norm. While antipreneurs come from a radically different side, they reject the need for change and offer resistance.
Since its introduction, the concept of antipreneurship has been tested and supplemented by empirical work (Bloomfield and Scott, 2018). These case-related studies often revolve around the analysis of nation-states in their role as antipreneurs in international negotiations and policy fields such as global security and human rights. Thereby, it is shown how antipreneurs neglect the claims of entrepreneurs that the status quo produces morally problematic outcomes. Certain tactical advantages for antipreneurs regularly result from heavily institutionalized norms that are difficult to change. And even if entrepreneurs succeed in justifying a problem that requires a solution, antipreneurs can move on to argue for maintaining the status quo rather than entering a phase of new uncertainty (Bloomfield, 2016; Bloomfield, 2018).
These findings are followed by additional empirical studies that analyze in detail the tactical maneuvers of antipreneurs. Lantis (2018: 209) shows how antipreneurs delay and even block significant progress in international forums. Similarly, Adachi (2018: 41) reveals tactics of diverting negotiations to consensus-driven institutions. Here, antipreneurs block the progress of the debate by making use of their veto. Another maneuver of the antipreneurs is to stress the importance of the norm of national sovereignty, whereby nation-states dismiss the emergence of new international norms as interventionalist (Bloomfield, 2018). Besides the reference to national sovereignty, Lantis and Bloomberg (2018: 165-166) add that antipreneurs also look for ways to have their justifications bolstered by protagonists who are on their side in the conflict over norms.
Thus, the concept of antipreneurship might be very applicable to the research of this work, as it focuses on the activity of norm resisters who provide a counterweight to the demands of the entrepreneurs and reject the emergence of a new norm. Thereby, the concept serves “as a sort of conceptual corrective by drawing attention to the possibility that resistance may both manifest and, at times, be quite effective” (Scott, 2018: 136). In case of LAWS, the application of the concept results in examining antipreneurs that are resisting a ban on LAWS in order to avoid any further international restriction beyond the already existing norms.
However, it must be kept in mind that the relationship between LAWS and the status quo of existing norms is controversially discussed, as it was shown in the literature review. This controversy is due to the circumstance that the weapons systems under discussion are rather still in the development phase. Typically, other weapons were already in use before a respective ban was discussed or implemented. Moreover, as already mentioned earlier, the definition of what should be considered as LAWS remains controversial, as concepts like autonomy and artificial intelligence are still difficult to grasp. Thus, doubts remain about the relationship of these emerging weapons systems to the global norms that exist around international security, the conduct of war and humanitarianism.
Due to these circumstances, the question arises to which existing norms antipreneurs refer in this case and how they also may draw advantages from the situation. This thought further motivates the relevance of applying the concept to the subject of LAWS and results in the second research question of this thesis.
3.4 Reflection on antipreneurship
From this departure it is useful to include a further perspective in order to reflect on the empirical application of the concept of antipreneurship. Recent research points out that there is still a great need for clarification regarding the demarcations and relationships of the concept, since the focus on the bipolar analysis between antipreneurs and entrepreneurs neglects actors who may not be assigned to either pole (Clapton, 2018; Lantis and Bloomberg, 2018: 166). In this sense, antipreneurs have to be separated both from those entrepreneurs who call for a preemptive ban on LAWS and from possible competing entrepreneurs who advocate an alternative approach. It has to be kept in mind for the analysis that it may remain difficult to clearly distinguish the thresholds from ambivalent behavior to open resistance and vice versa (Zahava, 2018).
Besides these definitional challenges, it is interesting to see how Bob (2018) suggests seeing conflict over norms as one between rival networks, which are often mirrored in their tactics and motivations to such an extent that they simultaneously promote and defend their own normative positions. It is emphasized that both groups of actors may apply affirmative strategies in which each actor advances its own position and negative strategies, in which each attacks the ideas of the enemy (Bob, 2018: 74). Originated is this analytical approach in case studies about gun control and homosexual rights (Bob, 2012). However, it is also adopted to questions of global security policy. In this case, it is shown that a network of actors considers targeted killing as a violation of human rights, while a rival network praises the practice as an efficient solution in the fight against terrorism (Bob, 2018).
On the one side, the applicability of the rival networks approach to the recent case of LAWS is rather limited, since the cases studies so far are more likely to be investigations of long-term conflicts between advocacy groups. On the other side, the approach extends the understanding of resistance against a ban on LAWS in three useful ways. First, Bob's network-based approach suggests focusing more on the similarities in the behavioral patterns of various antipreneurs and how these could be understood as a loosely organized network which refer to each other and adopt resistance patterns from each other.
Second, resistance does not necessarily have to be understood as a purely defensive role (Bob, 2018: 86). Therefore, attention could be paid to the dynamics of the conflict and how antipreneurs try not only to distract and delay the negotiations, but also shape them in their own interest and adjust their tactics. Accordingly, Campbell-Verduyn (2018) stresses that the strength of the concept of antipreneurship is that, when seen as the sharp opposite of the concept of entrepreneurship, a spectrum of role dynamics between these extremes opens up. Based on research about the 2008 financial crisis, he finds out that actors played the role of creative resistance by admitting that a minimal change in the status quo is acceptable. But their ultimate goal was in line with that of antipreneurs because they might give in a little while they still essentially defend the status quo. Third, empirical research on antipreneurs has so far lacked a consistent analytical approach. The case studies frequently lack clarity in the description of data collection and analysis. In this respect, Bob (2018; 2012: 16-35) presents a framework for the analysis of conflicts over norms. This framework can serve as a helpful guide to create a holistic approach to the analysis of antipreneurs. Consequently, the next chapter will present the research design that results from considerations presented in this chapter.
4 Research design
In the two previous chapters, relevant research was summarized from the two branches that are brought together in this thesis: the heterogeneous field of research on the topic of LAWS on the one hand and conceptual approaches regarding norms on the other hand. These two chapters clarified terminologies and offered important insights for the research design of this thesis. In the following, it is important to ensure a clear description of the research design so that the findings of the analysis can be traced by other researchers.
4.1 Scope of the case study
With regard to the existing empirical research on norm resistance, a trend towards the use of case studies, which typically focus on a specific subject area, has been identified. This established standard is continued in this thesis. The approach also appears valuable, since case studies help to examine the relationships and practices of actors in detail (Yin, 1994). However, the close analysis of a particular case also brings with it a limited degree of generalization. And the way the case is analyzed from a particular perspective leaves room for alternative interpretations based on other theoretical viewpoints. Herein lies the importance to clearly outline the scope, material and analysis of the case and to compare the findings of the own case study with already existing insights about antipreneurs. This case study will focus on the international meetings that have been conducted since 2014 under the auspices of the CCW. While in the first three years this took place at an informal basis, it was decided to initiate the GGE LAWS from 2017 onwards, whereas the GGE serves as the expert subsidiary body of the CCW. The eight meetings that have taken place so far between 2014 and 2019 will therefore form the scope of the analysis. Certainly, it may be criticized that the sole focus on the international discussions in the CCW forum limits the scope of the data introduced into the case study. However, the CCW is currently the only international forum in which specific talks about LAWS take place. The focus on this forum also facilitates the comparability of actors, since they all operate in the same institutional environment. For it may be assumed that agency can be limited by the surrounding institutional structures (Wendt, 1987).
The actors in this forum are in particular governmental delegations of nation-states. The case study will examine in detail the contributions of eight nation-states that are assumed to resist a ban on LAWS in the role of antipreneurs. Why the choice of several antipreneurs? This choice is based on the preceding idea that the eight antipreneurs could
be considered as a loosely connected network. The analysis is therefore particularly interested in exploring resistance patterns which the antipreneurs have in common and which are thus characteristic of their antipreneurial behavior.
The five permanent members of the United Nations Security Council can be regarded as most critical nation-states in terms of security policy and arms control (Stimmer, 2019). For this reason, the analysis focuses on the contributions of China, France, Russia, the UK and the USA, also because their institutional position may give much weight to their normative interpretations. Moreover, the literature review has shown that these countries have advanced capabilities in developing LAWS. One might argue that this narrowing would prevent the inclusion of interesting additional actors and that these five nation-states are too similar in terms of their strong power position and capabilities. Therefore, the analysis of antipreneurs also includes Australia, Israel and South Korea, as it is indicated that they also raise a voice against a ban in the GGE LAWS (Campaign to Stop Killer Robots, 2020). Israel and South Korea in particular are seen as being at the forefront of developing LAWS, as mentioned in the literature review, and may have a strategic interest in possessing them in relation to their conflictual surroundings.
Consequently, these are the eight actors which are assumed to be antipreneurs. As a point of comparison, three actors are included which may serve as important entrepreneurs of a ban: Brazil which is regarded as a major voice of the global South in this forum; Austria, as the only country of the European Union declared so far to support a ban on LAWS (Campaign to Stop Killer Robots, 2020); and the Campaign to Stop Killer Robots as the main non-governmental voice participating in the meetings.
Finally, the specific focus on these eleven protagonists serves two purposes. First, it is a diverse selection of actors in the forum who represent the opposing sides of the conflict over norms and include more and less militarily powerful states as well as nation-states from the global South and a non-governmental coalition. Second, the limitation to a sample of eleven actors will facilitate an in-depth analysis and a good cross-section of the relevant actors, while also acknowledging that other notable protagonists and their contributions must be excluded due to capacity reasons.
4.2 Collection of the data
Research about norms regularly lives from the use of qualitative, relatively unstructured data. This is also the case in order to gather data from the international meetings about LAWS. In the heart will be the data gathered from the contributions of the eleven actors,
which are under specific consideration. All of these actors submitted written statements and working papers to the eight sessions, which all took place under the auspices of the CCW in Geneva. These are the original documents of the governmental delegations of the ten nation-states and of the Campaign to Stop Killer Robots. The documents are accessible via the website of the CCW and in all cases, official English translations of the texts are used to facilitate the analysis.1
Besides, contextual data is included in the analysis to give credit to possible additional contributions and developments during these meetings apart from the eleven actors under investigation. That is why the written final reports and official agendas of the sessions, which are also accessible via the official website of the CCW, will be used as a source for data in order to review the general discussion lines and outcomes of each session. Particularly two points of criticism could be made with regard to the data collection. It could be argued that the case study should also include additional data sources: informal statements in the form of television interviews and tweets, for instance, or the gathering of interviews from persons participating in these sessions. However, this data is not collected because of its complexity of retrieval, transcription and linguistic diversity. It is assumed that this major expense would not be in line with its possible return, especially considering the limited time resources of the thesis project.
It could also be argued that data from civil society, media and domestic political actors should be included in the analysis. This criticism is acknowledged, and it is recognized that the analysis might be more in line with a “black box” mentality regarding the behavior of nation-states in international relations. However, the complexity of the search and the linguistic diversity of such possible data is at major expense of the time resources, as well. That is why such “domestic data” is not considered in this research process, also as it is only of minor interest when it comes to answering the research questions. While the inclusion of such data could be relevant in the course of further research projects, it is not to be considered essential in view of the aim of this thesis.
Finally, it is assumed that the collected textual data of the eleven actors together with the contextual material of the meetings may provide a sufficient basis for the upcoming analysis of the research questions.
1 The website with all collected documents of this case study is accessible here:
https://www.unog.ch/80256EE600585943/(httpPages)/8FA3C2562A60FF81C1257CE600393DF6?Open Document (last accessed:15/05/2020).
4.3 Analysis of the data
The challenge with all textual data is to use a transparent analytical approach so that the findings of the analysis are plausible. In empirical research about antipreneurs there is a tendency to present results without clearly describing the analytical approach, which reduces the transparency and reliability of data collection and analysis. This thesis aims to break away from this tendency.
Nevertheless, the design of a suitable approach for analyzing the resistance against a ban on LAWS is challenging for two reasons. First, there is no precedent for such a framework in empirical research on antipreneurship that could be applied to this case. Second, research suggests that resistance might be disclosed directly through counterarguments and justifications, which can be identified in the content of the statements and working papers of the antipreneurs. However, existing research also indicates that resistance might be shown in indirect forms of frustrating and obscuring the efforts of entrepreneurs and diverting negotiations (Bloomfield, 2018: 74). These tactics may not simply be found in the content of the texts, but through the interpretation of the tactical purposes that the statements contain in relation to the negotiating context (Bauer, 2011).
On the basis of these considerations, an analytical approach must be chosen that allows the exploration of resistance patterns through a close reading that moves between the content of the texts and their context. Here, the thesis draws on two suggestions to construct a framework for such an analysis. First, Bob's (2018; 2012: 16-35) framework for the analysis of rival networks serves as a helpful guideline as it poses guiding
questions to explore the development, dynamics and outcome of conflicts. Second, this
is linked to the idea that a useful strategy in the analysis of documents and their purposes may be to explore the ways in which things are said as potential solutions to problems (Gill, 2011: 178-181). In the context of the debate on LAWS, each antipreneur is faced with the problem of resisting the claim to ban LAWS. The analysis can therefore focus on the solutions that the antipreneurs commonly find to address this problem.
Accordingly, the analytical framework consists of questions that guide the discovery of resistance patterns which the antipreneurs have in common in solving the problem and which are thus characteristic of their antipreneurial behavior. The guiding questions are listed in figure 3 and result from the already existing findings sketched in the previous two chapters and which provide a first outline of solutions that antipreneurs might find for their resistance – especially with regard to the terminological hurdles regarding the definition of LAWS, the struggle over perceived problems and the presented findings of
empirical research on antipreneurship. These questions are guiding the analysis in that they form the analytical lenses through which the material is scanned in the following way:
Initially, all the statements and working papers of each antipreneur are reviewed. At this step, emphasis is placed on the guiding questions about direct resistance, as the contributions are particularly scrutinized in terms of how each antipreneur justifies resistance and contests a ban. These argumentative patterns of the antipreneurs are then compared with each other in order to consolidate their common characteristics. Then, the perspective is broadened by contrasting the texts of the antipreneurs with the context of the sessions. This context arises from the gathered contributions of the entrepreneurs as well as the reports and agendas of the sessions. Here, attention is paid to the questions of indirect resistance, as this step of the analysis aims to find out in what way the antipreneurs frustrate the efforts of the entrepreneurs and obstruct the discussions. In this manner, the patterns identified in the texts of the antipreneurs are specified and their purpose in terms of the discussions as well as delaying tactics become visible.
In the course of this gradual and systematic approach, the characteristic patterns with which the antipreneurs resist a preventive ban are identified, delimited and organized. Such a resistance pattern shall be recognizable in the behavior of all antipreneurs. It is conceivable that several patterns can be grouped, for example, because they build on each
other. A pattern could also consist of sub-patterns, because the antipreneurs display the same behavior but set different accents.
However, it must be taken into account that the eight nation-states are assumed to be antipreneurs. If it becomes apparent in the analysis that some of these nation-states deviate from the prevalent patterns of the majority of antipreneurs, the assumption that they behave as antipreneurs must be critically questioned in these cases. These deviations should also be made public, as this provides valuable information for the better delimitation of the concept of antipreneurship as demanded by critical research contributions (Clapton, 2018; Zahava, 2018).
One could argue that the analysis should not only be conducted on a qualitative basis, but that the quantification of patterns would also be helpful. This could give a numerical flavor to the analysis. However, it is noted that “validity is not a commodity that can be purchased with techniques […] Rather validity is like integrity, character and quality, to be assessed relative to purposes and circumstances” (Brinberg and McGrath, 1985: 13). Based on this quote, it is argued that it would be misleading to believe that quantification would add any benefit to the validity of the findings. For instance, resistance based on distraction or ignorance cannot be simply quantified. Rather, this analysis contributes to its validity by targeting the patterns that characterize the shared behavior of antipreneurs. Thus, it is argued that the common ground of consolidated resistance patterns provides a more adequate quality feature than seemingly simple attempts of quantification.
In addition to questions of validity, the limited generalizability of case studies must be taken into consideration. Thereby, Potter (1996) argues that each new study is a review of the adequacy of previous studies, what is particularly relevant in a case study dominated research area. Such case studies should “aim for a description of practice that is ‘rich’ enough to allow the research audience to judge its applicability to other contexts” (Yin, 1994: 43-44). This challenge is addressed through an analysis within a defined framework and the comparison with earlier findings. In this way, the thesis may serve to contrast and refine these existing insights about antipreneurship.
At the same time, it has to be kept in mind that this analysis remains an interpretation. Since it is never claimed that this is the only possible way to understand the texts or to specify the patterns, the present analysis is also an invitation to reinterpretation — the more so since all documents are publicly available. This fact, together with the transparently described research approach, strengthens the reliability of the findings that will be presented in the following.
5 Findings of the analysis
The previous chapters prepared the ground for the case study on antipreneurs in the international meetings on LAWS under the auspices of the CCW. In line with the two research questions of this thesis, the analysis aims to discover the patterns that the antipreneurs share in their resistance against a preventive ban on LAWS and the role that existing norms play in these patterns.
Subsequently, this analysis indicates that six of the eight nation-states that were originally assumed to be antipreneurs show resistance patterns that can be traced back to the concept of antipreneurship: Australia, Israel, Russia, South Korea, UK and USA. On the basis of the analytical framework outlined above and a close reading that moves between text and context, the analysis of the case was able to derive a total of eight resistance patterns which these six antipreneurs have in common.
It is also interpreted that these eight resistance patterns can be divided into groups. Two of them contribute in particular to block formal negotiations on a ban. Three other resistance patterns help to deny a problem between LAWS and the principles of IHL. And the last three patterns create a beneficial image about the emergence and development of LAWS. In addition, the eight resistance patterns can each be divided into a number of sub-patterns that highlight certain aspects of behavior. This structure of the findings – eight patterns that can be grouped as well as sub-divided – is shown in figure 4.
These findings will be presented over the course of the next three sub-chapters. In the presentation of the findings, several quotations from the analyzed documents are used and the findings are discussed close to the material. Additionally, a selection of text examples for each resistance pattern is shown in the annex of this thesis, see chapter 8, in order to exemplify which text passages are considered to represent these resistance patterns.2
After presenting the patterns which are characteristic to the behavior of the six antipreneurs, the analysis will turn to China and France. These two nation-states were assumed to be antipreneurs but their behavior deviates from the resistance patterns of the six antipreneurs. This is also an important finding, as it distinguishes the resistance of the antipreneurs from ambivalent behavior that cannot be clearly linked neither to the behavior of the entrepreneurs nor to that of the antipreneurs. Consequently, these findings of deviation are discussed in the fourth and final sub-chapter of the analysis.
5.1 Blocking formal negotiations on a ban
Consultations on LAWS under the auspices of the CCW started informally in 2014 and were further formalized with the establishment of the open-ended GGE LAWS in 2017. At first sight, entrepreneurs thus achieved the objective of rapidly placing LAWS on the institutional agenda (Rosert, 2019: 1109). At second glance, however, this development can also be interpreted as a great advantage for the activities of the antipreneurs because the GGE LAWS is based on full consensus. A forum based on consensus is best suited to block decisions, which is resulting in “effectively institutionalising the antipreneurs´ tactical advantages” (Adachi, 2018: 41).
These tactical advantages are particularly reflected in frustrating the progress of the sessions. This interpretation can be further underlined with a reference to the similar case of cluster munitions. Also, in this case the antipreneurs reacted very quickly to the campaigns by entrepreneurs. They took up the issue in the CCW and preemptively resisted before public support to ban the use of cluster munitions became even stronger and in order to regain control over the debate (Adachi, 2018: 44). The following two resistance patterns, which were discovered in the data, illustrate how this characteristic behavior of antipreneurs is also apparent in the case of LAWS.
5.1.1 Pattern A: Lack of definitions
The analysis indicates that antipreneurs resist a ban on LAWS by impeding the creation of an internationally accepted, common definition. Admittedly, antipreneurs contribute to the discussion about how LAWS are to be understood. Indeed, Russia, UK and the USA already provided their own definitions of LAWS. However, the wording of these definitions differs considerably from one another and reflects the problem of clarification outlined in chapter 2.1. The only point on which the antipreneurs largely agree is what should not be considered as LAWS: including unmanned aerial vehicles and other automated weapons systems that already exist. Thereby, they usually remind the parties “to remember that the CCW has been tasked with considering emerging technologies” (UK, 2016: 2).
Accordingly, the analysis shows that antipreneurs contribute to the sessions, but they do not make any effort to harmonize their definitions in order to enable a uniform understanding of autonomy. The lack of a common understanding frustrates the efforts of the entrepreneurs, who constantly challenge the forum “to agree on a carefully crafted and well balanced working definition of what LAWS actually are, so that a red line can
